Indonesian Journal of International Law (2021), Vol. 18 No. 3 pp.

423-446
https://doi.org/10.17304/ijil.vol18.3.818

AUTONOMOUS SPACE OBJECTS AND INTERNATIONAL

SPACE LAW: NAVIGATING THE LIABILITY GAP

Ioana Bratu, Arno R. Lodder & Tina van der Linden

Vrije Universiteit Amsterdam, the Netherlands

Correspondence: i.bratu@vu.nl

Abstract

The introduction of advanced new technologies is transforming the space industry. Artificial
intelligence is offering unprecedented possibilities for space-related activities because it enables
space objects to gain autonomy. The increasing autonomy level of space objects does not come
without legal implications. The lack of human control challenges existing liability frameworks.
This paper reviews the provisions of the Outer Space Treaty and the Liability Convention as the
main legal documents introducing the legal grounds for attributing liability in case of damages
caused by autonomous space objects. Looking at the limitations of these legal frameworks in
what concerns the attribution of liability, this paper identifies the conditions that could cause
a liability gap. The amendment of the Liability Convention, the concept of “international
responsibility” introduced by Article VI of the Outer Space Treaty and several international law
principles are analysed as potential solutions for preventing the liability gap and mitigating the
risks posed by autonomous space objects.

Keywords :artificial intelligence, autonomous space object, liability, responsibility, Liability

Convention, Outer Space Treaty

Submitted : 22 January 2021 | Revised : 14 February 2021 | Accepted : 30 March 2021

I. INTRODUCTION
Ever since Gagarin entered space, and a little later, Armstrong set foot on
the moon, governments spent large amounts of money on space-related activi-
ties. On a yearly basis, the global space economy is estimated by the Space
Foundation, an organization actively advocating on behalf of the global space
community, and by the Satellite Industry Association, engaged in advocating
on behalf of the commercial satellite industry of United States (US). Accord-
ing to these two sources, the global space economy is gradually increasing.
In 2018, the worth of the global space economy was estimated at $360 billion
(according to the Space Foundation) and approximately $415 billion (accord-
ing to the Satellite Industry Association).1
The increasing investments in the space industry also come from the pri-
1
“ESPI Yearbook 2019: Space Policies, Issues and Trends,” European Space Policy Institute,
https://espi.or.at/?view=article&id=468:espi-yearbook-2019&catid=29.

Copyright © 2021 – Bratu, Lodder & van der Linden

Published by Lembaga Pengkajian Hukum Internasional
Electronic copy available at: https://ssrn.com/abstract=3880911
Bratu, Lodder & van der Linden

vate sector, the so-called “New Space”, with major players such as SpaceX
and Blue Origin. This new phenomenon includes the emerging trends from the
space private business, which aims to engage in space-related activities inde-
pendently from governments. New entrants in the space industry usually fall
in one of two categories: the first is existing large companies, such as Google
or Facebook, interested in diversifying their portfolio and creating a symbiosis
between their current business activities and space applications and the sec-
ond is new space companies: start-ups.2 In terms of space budgets allocated
by the private space sector, the last decade showed a significant increase in
investments. Over the last 15 years, the total investment in space-related start-
up ventures amounts to $13.3 billion and more than 80 new space companies
have been set up.3
In addition to being the new actors, New Space also includes innovative
industrial approaches, specifically in what concerns advanced new technolo-
gies. Space-related technologies contributed to the growth of the private sec-
tor and by developing innovative technologies will continue to do so in the
near future, as indicated by Jet Propulsion Laboratory (JPL), a global leader in
planetary exploration and space-based astronomy that supports the missions
of National Aeronautics and Space Administration (NASA). Regarding these
innovative technologies, developing autonomous systems is a top priority.4
Autonomous systems are equipped with artificial intelligence (AI) capabilities
and function without human intervention.
The introduction of autonomous systems in space activities does not come
without legal implications, especially in what concerns issues of liability. As
far as we know there are no cases yet requiring the application of space law
in the context of damages caused by an AI system. However, if we look at
the evolution of AI systems in other industries, such as the self-driving cars
industry, many incidents already occurred.5 We believe we should anticipate
future incidents involving autonomous space objects and consider a frame-
work for liability regimes, in order to avoid situations in which liability cannot
be attributed, in other words: a liability gap. Therefore, this paper analyses
whether existing legal frameworks dealing with liability for damages caused
by space objects are capable of dealing with incidents caused by AI, specifi-

2
Ibid.
3
Alessandra Vernile, The Rise of Private Actors in the Space Sector (Springer International
Publishing, 2018), 2.
4
Strategic Technologies: Science and Technology, Jet Propulsion Laboratory (California
Institute of Technology, 2019), 2.
5
“Self-Driving Car Statistics for 2021: Policy Advice,” accessed February 14, 2021, https://
policyadvice.net/insurance/insights/self-driving-car-statistics/.

424

Electronic copy available at: https://ssrn.com/abstract=3880911

 Autonomous Space Objects and International Space Law

cally, by autonomous space objects.

II. ARTIFICIAL INTELLIGENCE FOR SPACE ACTIVITIES

Defining AI is no easy task. The concept itself is broad and different ap-
proaches have proposed different definitions.6 Despite the increased interest in
AI by academia, industry and public institutions, there is no standard defini-
tion of AI systems. Mostly, AI is described by reference to the historical evo-
lution of this concept, its corresponding evolution of capabilities, and the use
of technology that performs tasks requiring human intelligence.7

A. THE CONCEPT OF ARTIFICIAL INTELLIGENCE

In this paper, the definition of AI draws on the European Commission’s
reports, the expert groups appointed by the European Commission and by
the European Parliament. In accordance with the opinion of the Independent
High-Level Expert Group on Artificial Intelligence, AI systems aim to real-
ize a goal, by acting in the physical or digital dimension by perceiving their
environment through data acquisition, interpreting the collected structured or
unstructured data, reasoning on the knowledge, or processing the information,
derived from this data and deciding on the best action(s) to take to achieve
the given goal.8 Pursuant to European Parliament’s Committee on Legal Af-
fairs, in some cases, AI systems can adapt their behaviour by analysing how
the environment is affected by their previous actions.9 In a different position
paper, the European Commission states that AI refers to systems that display
intelligent behaviour by analysing their environment and taking appropriate
action to achieve specific goals with some degree of autonomy.10
AI is already used in various fields, such as healthcare, transport, finance,

6
Virginia Dignum, “What Is Artificial Intelligence?” in Responsible Artificial Intelligence:
How to Develop and Use AI in a Responsible Way, Virginia Dignum, ed. (Cham: Springer
International Publishing, 2019), 9–34.
7
Sofia Samoli et.al, “AI Watch: Defining Artificial Intelligence,” (Luxembourg: Publications
Office of the European Union, 2020).
8
European Commission, “A Definition of Artificial Intelligence: Main Capabilities and
Scientific Disciplines,” in Independent High-Level Expert Group on Artificial Intelligence set
up by the European Commission, 8 April 2019.
9
Nathalie Nevejans, European Civil Law Rules in Robotics (Brussels: European Parliament,
2016).
10
“Communication from the Commission to the European Parliament, the European Council,
the European Economic and Social Committee and the Committee of the Regions, Coordinated
Plan on Artificial Intelligence,” European Commission Brussels, 7 December 2018, https://eur-
lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:52018DC0795&from=EN.

425

Electronic copy available at: https://ssrn.com/abstract=3880911

Bratu, Lodder & van der Linden

personnel recruitment etc.11 AI systems can be descriptive, they tell you what
happened; diagnostic, they tell you why something happened or predictive as
they forecast what will (statistically probably) happen; and prescriptive in the
sense of being capable of performing actual decision making and implemen-
tation.12
The process of making decisions and taking actions of AI systems is ena-
bled by the fact that the system is fed with a relevant set of data or uses ap-
propriate sensors, for example, cameras or microphones, enabling the system
to collect the data required for achieving the goal for which it was designed.13
Subsequently, the collected data are interpreted and the system takes a deci-
sion, which may be translated into either pursuing an action or not. If it de-
cides to act, this decision will be executed through the system’s physical or
software actuators.14
Depending on the type of the AI system, the final decision is made either
by humans or autonomously, sometimes with some degree of human control.
AI systems can make decisions and improve their capabilities without human
intervention but depending on the available data. The process in which possi-
ble new actions are considered through an analysis of desired outcomes based
on previous failure or success is known as machine learning (ML).15 The in-
spiration for this comes from the neural networks of the human brain. As a
general classification, there are two main categories of ML: supervised and
unsupervised. Supervised ML relies on algorithms, which have been trained to
calculate outcomes based on examples, i.e. the AI system was “trained” with
examples of sets of input and corresponding output data previously identified
as correct.16 For unsupervised learning, algorithms are not trained, do not re-
ceive instructions identifying which data sets are correct and, therefore, will

11
“What Is Artificial Intelligence and How Is It Used?” News European Parliament, 9 April
2020, https://www.europarl.europa.eu/news/en/headlines/society/20200827STO85804/what-
is-artificial-intelligence-and-how-is-it-used.
12
Humberto Farias, “Machine Learning vs. Predictive Analytics: What’s the Difference?”
Concepta, 10 October 2017, accessed 11 February 2021, https://www.conceptatech.com/blog/
machine-learning-vs-predictive-analytics-what-is-the-difference.
13
Basheer Qolomany, et. al., “Leveraging Machine Learning and Big Data for Smart Buildings:
A Comprehensive Survey,” IEEE Access 7, (2019): 90316–90356.
14
“How Artificial Intelligence Works,” European Parliament Think Tank, 14 March 2019,
accessed 14 February 2021, https://www.europarl.europa.eu/thinktank/en/document.
html?reference=EPRS_BRI(2019)634420.
15
“Machine Learning: What It Is and Why It Matters,” SAS, accessed 14 February 2021,
https://www.sas.com/en_us/insights/analytics/machine-learning.html.
16
Stuart Russell and Peter Norvig, Artificial Intelligence: A Modern Approach (Third edition)
(Harlow: Pearson, 2014), 695.

426

Electronic copy available at: https://ssrn.com/abstract=3880911

 Autonomous Space Objects and International Space Law

search independently for relevant data sets required for achieving their goal.17
In the space industry, one area in which the applications of AI are cur-
rently being investigated is in satellite operations; in particular, to support the
operations of large satellite constellations, such as positioning, communica-
tion and end-of-life management. In addition, it is becoming more common
to find ML systems analysing the huge amount of data that comes from each
space mission.18

B. PRACTICAL APPLICATIONS OF ARTIFICIAL INTELLI-

GENCE FOR SPACE ACTIVITIES
Using AI for space-related activities enable space objects to gain auton-
omy. This offers a series of advantages, such as spacecraft being capable to
rapidly assess and react to events and changing environments, thus increasing
the reliability and productivity of missions.19 In the absence of AI capabilities,
spacecraft cannot determine their own operational status and make decisions
on their own. In space missions dependent on the ground segment, many hu-
man experts are required, making such missions cumbersome and time-con-
suming. Sometimes, it may take days before the data are processed, decisions
are made, and uploaded commands reach the spacecraft.20 According to the
European Space Agency (ESA) the AI disruption is emphasised by its con-
vergence with other transformative technologies such as IoT, cloud comput-
ing and blockchain, which are transforming entire industry verticals such as
automotive, healthcare, transport and banking. For example, the emergence
of self-driving cars, is made possible today due to the convergence and inte-
gration of technologies such as IoT, cloud computing and AI. Will a similar
disruption occur in the space sector? Will AI be the key to unlock the potential
of the new streams of Earth Observation (EO) data coming online to better
understand changes on Earth? Will satellite hardware become a commodity
focusing on the AI-powered software enabling autonomy and remote upgrade,
as is happening with Tesla cars? In this context, one of the key challenges for
the EO community is to be able to exploit the full power of AI in collaboration
with new players in the ecosystem including ICT companies, start-ups and
17
“Big Data, Artificial Intelligence, Machine Learning and Data Protection,” Information
Commissioner’s Office, 2017, https://ico.org.uk/media/for-organisations/documents/2013559/
big-data-ai-ml-and-data-protection.pdf.
18
European Space Agency, “Robots in Space,” accessed 12 February 2021, https://www.esa.int/
Enabling_Support/Preparing_for_the_Future/Discovery_and_Preparation/Robots_in_space2
19
Ibid.
20
Daniela Girimonte and Dario Izzo, “Artificial Intelligence for Space Applications,” in
Intelligent Computing Everywhere, Alfons J. Schuster, ed. (London: Springer London, 2007),
235-253.

427

Electronic copy available at: https://ssrn.com/abstract=3880911

Bratu, Lodder & van der Linden

data and EO scientists.21

There is a growing necessity for processing the extensive amount of data
generated by the new generation of satellites, such as the Copernicus pro-
gramme’s data.22 Moreover, the success of future deep space missions (e.g.,
travelling to Mars) will be dependent on various types of AI applications.23
Also, by enabling onboard decision-making, AI technologies support space-
craft in detecting events. Using AI maybe even access to dynamic environ-
ments, such as comets, will become feasible.24
AI is also used to monitor the operation of satellites by watching the pat-
terns of other satellites, planets, and space debris and carrying out correc-
tive actions if needed. Such monitoring capabilities are used, for example,
by SpaceX for avoiding collision of satellites with other objects.25 Another
recent example relates to space station operations. The Crew Interactive Mo-
bile Companion (CIMON) is the world’s first flying autonomous astronaut
assistant featuring AI,26 introduced on the International Space Station in 2018.
CIMON enables fully voice-controlled access to documents and media, and it
can conveniently navigate through operating and repair instructions and pro-
cedures for experiments and equipment.27 In terms of planetary exploration,
AI is used to navigate on-site conditions that are still too dangerous for hu-
mans. The rover Perseverance was launched by US for navigating on Mars.28
The technology behind it is similar to the one used by self-driving vehicles,
however, with the major difference, that this rover has to navigate more com-
plicated terrain, which is analysed by the computer vision systems installed in
the rover as it moves. If a terrain problem is encountered, the AI system takes
21
Pierre-Philippe Mathieu, Sveinung Loekken, et. al., “Towards a European AI for Earth
Observation Research & Innovation Agenda,” in Workshop at ESA Φ-lab (European Space
Agency, 2018), 1-20, https://blogs.esa.int/philab/files/2018/07/Towards-a-European-AI-for-
Earth-Observation-Research-Innovation-Agenda-.pdf.
22
Ibid.
23
European Space Agency, “Robots in Space.”
24
Steve Chien, et. al., “The Future of AI in Space,” IEEE Intelligent Systems 21, no. 4 (July 1,
2006): 64–69.
25
Ron Schmelzer, “How is AI Helping to Commercialize Space?” Forbes, 21 March 2020,
accessed February 12, 2021, https://www.forbes.com/sites/cognitiveworld/2020/03/21/how-is-
ai-helping-to-commercialize-space.
26
“Astronaut Assistant CIMON-2 Is on its way to the International Space Station,” Airbus,
accessed 12 February 2021, https://www.airbus.com/newsroom/press-releases/en/2019/12/
astronaut-assistant-cimon2-is-on-its-way-to-the-international-space-station.html.
27
“Floating Robot Cimon sent to International Space Station,” BBC News, 29 June 2018,
accessed February 12, 2021, https://www.bbc.com/news/technology-44655675.
28
Daniel Oberhaus, “How NASA Built a Self-driving Car for Its Next Mars Mission,” Wired,
21 July 2020, accessed February 12, 2021, https://www.wired.com/story/how-nasa-built-a-self-
driving-car-for-its-next-mars-mission/.

428

Electronic copy available at: https://ssrn.com/abstract=3880911

 Autonomous Space Objects and International Space Law

a decision and changes the course of the rover in order to avoid it.29 AI is also
used for space sustainability, i.e. for removing space junk. ESA plans to launch
the world’s first debris-removing space mission, ClearSpace-1, which will use
an AI-powered camera to find the debris. Its robotic arms will then grab the
object and drag it back to the atmosphere in order to burn it up.30 Another ap-
proach in using AI for space sustainability includes collision avoidance ma-
noeuvres using ML techniques.31 This collision avoidance system, currently
under development by ESA, will automatically assess the risk and likelihood
of in-space collisions, improve the decision making process on whether or not
a manoeuvre is needed, and may even send the orders to at-risk satellites to
get out of the way.32
These examples of how AI can be used in space are paving the way for a
higher autonomy level of this technology, which will be required to achieve
other important milestones in space-related activities, for example reaching
out to neighbouring solar systems, as Alpha Centauri. This would imply the
traversing of a distance of over four light-years. Upon arrival, the spacecraft
would need to operate independently for years, even decades, exploring mul-
tiple planets in the solar system. This ambition is not far from becoming a
reality, given recent precedents: for example, in 2017, an autonomous space-
craft completed almost a dozen years of nearly continuous operations of Earth
observation, using both onboard and ground-based AI.33

III. SPACE LAW LIABILITY REGIMES UNDER THE UN IN-

TERNATIONAL TREATIES
The space treaties form the core of the whole space legal system con-
cluded within the framework of the United Nations, by the Committee on the
Peaceful Uses of Outer Space. Out of these, the Outer Space Treaty34 and the
29
“NASA’s Mars Rover Drivers Need Your Help,” Jet Propulsion Laboratory, 12 June 2020,
accessed February 12, 2021, https://www.jpl.nasa.gov/news/nasas-mars-rover-drivers-need-
your-help/.
30
Thomas Macaulay, “AI to Help World’s First Removal of Space Debris,” Neural, 30 October
2020, accessed 12 February 2021, https://thenextweb.com/neural/2020/10/30/ai-to-help-
worlds-first-removal-of-space-debris.
31
Audrey Berquand and Deep Bandivadekar, “Five Ways Artificial Intelligence Can Help
Space Exploration,” The Conversation, 25 January 2021, accessed February 12, 2021, http://
theconversation.com/five-ways-artificial-intelligence-can-help-space-exploration-153664.
32
“Automating Collision Avoidance,” European Space Agency, accessed February 12, 2021,
https://www.esa.int/Safety_Security/Space_Debris/Automating_collision_avoidance.
33
Steve Chien and Kiri L. Wagstaff, “Robotic Space Exploration Agents,” in Science Robotics
2, no. 7 (2017): 1-2.
34
Treaty on Principles Governing the Activities of States in the Exploration and Use of Outer

429

Electronic copy available at: https://ssrn.com/abstract=3880911

Bratu, Lodder & van der Linden

Liability Convention35 provide the legal grounds for attributing liability in

case of damage caused by space objects. Article VII of the Outer Space Treaty
sets forth the principle of international liability of the launching state for dam-
age caused by its space object. However, several key-issues remained unad-
dressed, such as: what qualifies as damage, what type of liability regime is ap-
plicable or what procedure should be followed in case damage occurs.36 Given
these issues, the Liability Convention’s purpose was to elaborate on Article
VII of the Outer Space Treaty and to establish a detailed liability regime.
The characteristics of the Liability Convention together with the corre-
sponding provisions of the Outer Space Treaty are discussed below.

A. LIABILITY VERSUS RESPONSIBILITY

First of all, a distinction needs to be made between two concepts of ac-
countability, liability and responsibility, which are used in the same conven-
tion in an outer space context:
1. Article VI of the Outer Space Treaty provides that states shall bear
“international responsibility” for national activities in outer space,
2. Article VII of the same Treaty refers to “international liability” to be
attributed to a state that launches or procures the launching of an ob-
ject into outer space, in case damage is caused to another state or to
its natural or juridical persons by such object or its component parts.
The Liability Convention details the provisions of Article VII of the Outer
Space Treaty differentiating between absolute and fault-based liability, de-
pending on the location of the damages. Neither “liability” nor “responsibil-
ity” has been defined in the UN International Treaties. The term “liability” is
used to establish the launching state’s liability for damage caused by space
objects, while the term “responsibility” has been used to mandate internation-
al responsibility by the appropriate state party for national activities in outer
space.37 From this it may be concluded that “liabilities” are mentioned when
we are dealing with legal consequences, mostly in terms of damages, aris-
ing from a particular behaviour. On the opposite side, it seems that when we
speak of responsibilities, we are dealing primarily with obligations imposed
on people and institutions who are supposed to carry out certain activities

Space, Including the Moon and Other Celestial Bodies, 610 U.N.T.S. 205 [Outer Space Treaty].
35
United Nations Convention on the International Liability for damage caused by space objects,
opened for signature on 29 March 1972, 961 UNTS 13810 (entered into force September 1972).
36
Fabio Tronchetti, Fundamentals of Space Law and Policy (New York: Springer, 2013), 72.
37
Bin Cheng, “International Responsibility and Liability for Launch Activities,” Air and Space
Law 20, no. 6 (December 1, 1995): 297-310.

430

Electronic copy available at: https://ssrn.com/abstract=3880911

 Autonomous Space Objects and International Space Law

or are accountable in given situations though not necessarily in the form of

compensation for damages. Therefore, it appears that to some extent the two
concepts are interrelated.38
An in-depth analysis of the potential confusion between the two concepts
of accountability is beyond the purpose of this paper.39 However, given the
challenges posed by the rapid introduction of advanced AI systems in space-
related activities and the risks related to attributing liability under the Liability
Convention, it is worth mentioning that there is no fundamental reason why
the more general concept of state responsibility could not be used also for
obtaining compensation for damage in cases where the liability concept may
not offer this possibility. Thus, at least in theory, a state that is the victim
of damage caused by an unlawful act might choose to seek compensation
for such injuries not from the state technically liable under Article VII of the

Outer Space Treaty and the corresponding Articles II and III of the Liability
Convention, but from the state technically responsible under Article VI of the
Outer Space Treaty.40
Figure 1. Accountability Concepts under Core International Space Law

B. ABSOLUTE AND FAULT-BASED LIABILITY

Under the Liability Convention, liability is by definition attributed only
38
Stephen Gorove, “Liability in Space Law: An Overview Space Law,” Annals of Air and
Space Law 8, (1983): 373–380.
39
See for this issue in detail, Frans von der Dunk “Liability versus Responsibility in Space
Law: Misconception or Misconstruction?” in Proceedings of the Thirty-Fourth Colloquium on
the Law of Outer Space, 1992, 363–371.
40
Frans von der Dunk, “International Space Law,” in Handbook of Space Law, Frans von der
Dunk and Fabio Fonchetti, eds. (Cheltenham, United Kingdom: Edward Elgar Publishing), 50.

431

Electronic copy available at: https://ssrn.com/abstract=3880911

Bratu, Lodder & van der Linden

to states, specifically to “launching states”, which include: (i) a state which

launches or procures the launching of a space object or (ii) a state from whose
territory or facility a space object is launched.41 The Convention distinguishes
between the liability of states related to damage caused by its space object
on the surface of the Earth or to aircraft in flight, which is absolute42 and to
damage caused not on the surface of the Earth.43 For the latter type of damage
the state can be held liable only if the damage is due to its fault or the fault
of persons for whom it is responsible. Thus, the geographical location of the
damage is a fundamental criterium44 for attributing liability, i.e., on the surface
of the Earth or elsewhere.
1. Absolute Liability
The reason behind the Liability Convention’s introduction of absolute lia-
bility is the fact that space activities create an extraordinary risk to persons and
property with the supplementary inconvenience of the difficulty of establish-
ing the proof of fault in the case of an accident caused by a space object. In or-
der to protect victims, the burden of proof was therefore put on the launching
state, since it supposedly benefits from the activity which caused the damage.
It is also important to note that it is not a general principle of international law
that states can be held absolutely liable, but rather an exceptional situation. As
a rule, the establishment of fault is required before liability can be attributed.45
Thus, it can be argued that the Liability Convention has created a new situa-
tion which so far has not been reproduced by any other international treaty.46
2. Fault Liability
The reason behind the introduction of a fault liability regime is that, in
space, all parties in the position to operate a space object are assumed to be
acting on an equal footing, to have the technology to provide the proof of the
fault, and in any case to have assumed the risks of conducting these activities:
none of them should be a “privileged victim”.47
The notion of “fault”, or more precisely, the interpretation of this notion,
received a lot of attention in the legal doctrine because the Liability Conven-
tion does not define the term. The Convention also fails to refer to a duty of
41
Frans von der Dunk, Advanced Introduction to Space Law (Cheltenham, United Kingdom:
Edward Elgar Publishing, 2020), 49.
42
Liability Convention, art II.
43
liability Convention, art III.
44
Valérie Kayser, Launching Space Objects:  Issues of Liability and Future Prospects
(Dordrecht: Springer, 2001), 50.
45
Ibid.
46
von der Dunk and Tronchetti, Handbook of Space Law, 89.
47
Valérie Kayser, Launching Space Objects, 51.

432

Electronic copy available at: https://ssrn.com/abstract=3880911

 Autonomous Space Objects and International Space Law

care in outer space, the breach of which would constitute fault.48 As a general
rule, fault can only be asserted when the act violates duty of care or standard
of conduct. The Convention does not provide a clear obligation to act or to
abstain from acting, making it difficult to implement this notion in practice.49
Some legal scholars referred to Black’s Law Dictionary in an attempt to define
fault as “an error or defect of judgement or of conduct; any deviation from
prudence or duty resulting from inattention (…); the intentional or negligent
failure to maintain some standard of conduct when the failure results in harm
to another person”.50 As such, “fault liability” can then be defined as imply-
ing a certain degree of blameworthiness, or, alternatively, a type of liability
in which the plaintiff must prove that the defendant’s conduct was either neg-
ligent or intentional.51 Other legal scholars considered that the Liability Con-
vention did not explicitly resolve certain details on purpose, for instance de-
fining “fault”, in order to avoid a too specific approach. If the negotiations of
the Liability Convention had moved in that direction most likely the Liability
Convention would have never been agreed upon. Fortunately, it was possible
to agree on a formal process for the resolution of disputes, therefore, some of
the lacunae in the Convention can be resolved through the use of this process.
Thus, it may, for example, be possible to obtain the required understanding of
the meaning of “fault,” if necessary.52

IV. AUTONOMOUS SPACE OBJECTS AND CHALLENGES

POSED TO SPACE LAW LIABILITY REGIMES
In the near future, most likely human intervention will still be required be-
yond the initial programming of the AI system, which would entail that states
remain liable for national space activities and space objects equipped with AI
capabilities. However, the hypothetical case when a state deploys in space a
fully autonomous space object raises important questions due to the removal

48
Marc S. Firestone, “Problems in the Resolution of Disputes Concerning Damage Caused
in Outer Space,” Tulane Law Review : Devoted to the Civil Law, Comparative Law and
Codification, 1985.
49
Yun Zhao, “The 1972 Liability Convention: Time for Revision?” Space Policy 20, no. 2
(May 1, 2004): 117–22, https://doi.org/10.1016/j.spacepol.2004.02.008.
50
Frans von der Dunk, “Too-Close Encounters of the Third Party Kind: Will the Liability
Convention Stand the Test of the Cosmos 2251-Iridium 33 Collision?,” Space, Cyber,
and Telecommunications Law Program Faculty Publications, January 1, 2010, https://
digitalcommons.unl.edu/spacelaw/28.
51
Ibid.
52
Carl Q. Christol, ““International Liability for Damage Caused by Space Objects”, American
Journal Of International Law 74, no. 2 (1980): 346-371, doi:10.2307/2201505.

433

Electronic copy available at: https://ssrn.com/abstract=3880911

Bratu, Lodder & van der Linden

of human judgment from the equation.53 In case of incidents involving autono-

mous space objects, significant liability issues may arise in the event of, for
example, collisions or destructions. Therefore, the use of AI in space-related
activities require revisiting the traditional concepts of liability under the Outer
Space Treaty and the Liability Convention.

A. THE NOTION OF “SPACE OBJECT”

A first step in the process of attributing liability would require investigat-
ing if the provisions of the UN International Treaties are sufficient to deal
resolve any claims arising out of incidents involving autonomous space ob-
jects. The Liability Convention does not provide a clear definition for the term
“space objects”. It only mentions that “the term space object includes compo-
nent parts of a space object as well as its launch vehicle and parts thereof”.54
Taking into consideration the limited technologies available when the
Convention was drafted, the underlying assumption may have been that
a launch constituted a vertical departure from a land-based launch facility
aimed at entering outer space using rocket engines. The technological devel-
opments challenge the existing notion of “launch vehicle” and consequently,
the entire notion of “space object”. New technologies already started to fa-
cilitate air launches55 and, more recently, autonomous air launches.56 In this
context, the definition of a “space object” should be interpreted as including
any man-made object which is attempted to be physically brought into outer
space.57 Moreover, any piece of hardware used in a launch together with all
other pieces collectively constitute a space object, and states cannot choose
what does, or does not, make up the object for the purposes of this legal defi-
nition.58 Based on this interpretation, the term “space object” would also need
to include any AI software and any technical features enabling autonomy of
a space object.
Autonomous space objects require extra attention also in what concerns

53
Anne-Sophie Martin and Steven Freeland, “The Advent of Artificial Intelligence in Space
Activities: New Legal Challenges,” Space Policy 55 (February 1, 2021): 101408, https://doi.
org/10.1016/j.spacepol.2020.101408.
54
Liability Convention, art I (b).
55
von der Dunk and Tronchetti, Handbook of Space Law, 86.
56
Eric Berger, “Meet Ravn X—A fully autonomous, air-launched rocket for small satellites,”
Ars Technica, accessed 20 December 2020, https://arstechnica.com/science/2020/12/meet-
ravn-x-a-fully-autonomous-air-launched-rocket-for-small-satellites/.
57
Vladimir Kopal, “Some Remarks on Issues Relating to Legal Definitions of Space Objects,
Space Debris and Astronaut,” in Proceedings of the 37th on the Law of Outer Space (1994), 99.
58
Stephan Hobe, Cologne Commentary on Space Law / Vol. 2, Rescue Agreement, Liability
Convention, Registration Convention, Moon Agreement. (Köln: Heymann, 2013), 34.

434

Electronic copy available at: https://ssrn.com/abstract=3880911

 Autonomous Space Objects and International Space Law

registration requirements under the UN Registration Convention.59 Registra-

tion of space objects is an indispensable means for administering good care of
space objects relating directly to legal issues such as jurisdiction and control
over space objects.60 The Registration Convention provides that when a space
object is launched into Earth orbit or beyond, the launching State shall register
the space object by means of an entry in an appropriate registry which it shall
maintain; where there are two or more launching States in respect of any such
space object, they shall jointly determine which one of them shall register the
object.61 The Registration Convention also provides that each State of regis-
try shall furnish to the Secretary-General of the United Nations, as soon as
practicable, the following information concerning each space object carried
on its registry: name of launching State or States; an appropriate designator
of the space object or its registration number; date and territory or location
of launch; basic orbital parameters; general function of the space object.62
Given the specific capabilities of autonomous space objects, some authors
recommend taking in consideration the possibility of developing a ‘special’
registry that specifies the unique features of spacecraft having on board AI
capabilities.63 It is also duly noted that this suggestion will require additional
investigations to see whether, and if so how, this should be implemented.64

B. THE NOTION OF “GROSS-NEGLIGENCE”

In what concerns the absolute liability described under Article II of the Li-
ability Convention, we do not envisage any particular difficulties in relation to
attributing this type of liability. States are to be held absolutely liable for the
damages caused by their space objects on the surface of the Earth irrespective
of their autonomous capabilities. The situation is more complicated in case of
the exoneration procedure described under Article VI of the Liability Conven-
tion: “exoneration from absolute liability shall be granted to the extent that
a launching State establishes that the damage has resulted either wholly or
partially from gross negligence or from an act or omission done with intent to
cause damage on the part of a claimant State or of natural or juridical persons
it represents”. The provisions of this article relate to the exoneration of liabil-
ity for damages caused by the claimant state’s own gross negligence subject

59
United Nations Convention on Registration of Objects Launched into Outer Space, Nov. 12,
1975, 28 U.S.T. 695, 1023 U.N.T.S. 15. [Registration Convention]
60
Yoon Lee, “Registration of Space Objects: ESA Member States’ Practice,” Space Policy 22,
no. 1 (February 1, 2006): 42–51, https://doi.org/10.1016/j.spacepol.2005.11.007.
61
Registration Convention, art. II
62
Registration Convention, art. IV
63
Martin and Freeland, “The Advent of Artificial Intelligence in Space Activities”
64
Ibid.

435

Electronic copy available at: https://ssrn.com/abstract=3880911

Bratu, Lodder & van der Linden

that the launching state had not violated international law and the international
UN treaties.65
The Liability Convention does not provide for a definition of “gross negli-
gence” and no indication was made for attributing negligent conduct to others
or for the allocation of a principal’s vicarious liability for an agent or employ-
ee.66 In the absence of clear criteria applicable to gross negligence, it may
prove difficult to apply them in practice. According to the Cambridge Diction-
ary, gross negligence is being defined as a serious lack of care or attention
towards a person or thing that another person is responsible for.67 The analysis
of this definition reveals that the concept is related to the mental element of an
act or omission, it is a product of human thought, it is associated to an action
or omission part of a human activity. The concept is being challenged in case
of a damage resulting wholly or partially from an act or omission of an au-
tonomous space object deployed or controlled by a claimant state. Depending
on the autonomy level of the space object, invoking exoneration of liability
based on the concept of gross negligence may not be applicable in case of
space objects equipped with AI capabilities,68 because it no longer involves a
human activity.

C. THE NOTIONS OF “FAULT” AND “PERSON”

The fault-based liability regime introduced by Article III of the Liability
Convention also raises similar problems in what concerns its applicability to
potential incidents involving autonomous space objects. The absence of a def-
inition concerning the notion of “fault” or clear criteria for assessing fault may
cause difficulties in practice.69 The same applies to the notion of “persons”,
which was not defined. The notion “person”, as it is used Article III generally
refers to an entity which is subject to legal rights and duties such as a natural
or juridical person. The law considers artificial entities like corporations, part-
nerships, joint ventures, and trusts to be “persons” as they are subject to legal
rights and duties. Additionally, in certain instances, the law recognizes and
imposes legal rights and duties on certain inanimate objects like ships, land,
and goods which results in such inanimate objects being subject to adjudica-

65
Liability Convention, art. VI (2)
66
Christol, “International Liability for Damage Caused by Space Objects”.
67
“Gross Negligence,” Cambridge Dictionary, accessed 20 December 2020, https://dictionary.
cambridge.org/dictionary/english/gross-negligence.
68
George Anthony Gal, Cristiana Santos, Lucien Rapp, Réka Markovich, Leendert van
der Torre, “Artificial intelligence in space,” available at https://www.researchgate.net/
publication/342377395_Artificial_intelligence_in_space.
69
See Supra 3.2, II

436

Electronic copy available at: https://ssrn.com/abstract=3880911

 Autonomous Space Objects and International Space Law

tive jurisdiction as well as being subject to a judgment rendered against it.70

Related to the case of autonomous space objects, i.e., space objects
equipped with AI capabilities, the discussions surrounding the possibility of
granting legal personality to AI systems received a lot of attention in academic
literature. In general, the analysis concerns the question of whether or not one
can argue that an AI system possesses the necessary capabilities to be consid-
ered full moral and legal persons.71 The topic was also discussed by one of
the expert groups of the European Commission, part of the legislative train
schedule related to the future of AI governance. Even if only formulated as a
recommendation, the report includes a straightforward message concerning
this topic, indicating that there is currently no need to give a legal personality
to emerging digital technologies, such as AI. This is because harm caused by
even fully autonomous technologies can generally be reduced to risks attrib-
utable to natural persons or existing categories of legal persons, and where
this is not the case, new laws directed at individuals are a better response than
creating a new category of legal person.72
Given the above, it may be concluded that, since fault liability under Ar-
ticle III of the Liability Convention is premised on the fault of a State or the
faults of persons, a decision by an autonomous space object will, in all likeli-
hood, not be the “fault of persons”.73 This may cause difficulties in attributing
liability based on the fault-based regime under the Liability Convention, thus
potentially leading to a liability gap.

V. PROPOSED SOLUTIONS FOR MITIGATING LIABILITY

GAPS
Several solutions can be proposed for mitigating the challenges posed by
the autonomous space objects to the liability regimes. Even if they have not
been tested in practice yet, these proposals are mentioned here for the purpose
of avoiding situations in which liability cannot be attributed, in other words a
liability gap.

70
Ibid.
71
David J. Gunkel, “The Other Question: Can and Should Robots Have Rights?,” Ethics and
Information Technology 20, no. 2 (2017): pp. 87-99, https://doi.org/10.1007/s10676-017-9442-
4.
72
Publications Office of the European Union, “Liability for Artificial Intelligence and other
Emerging Digital Technologies.” (Publications Office of the European Union, November
27, 2019), http://op.europa.eu/en/publication-detail/-/publication/1c5e30be-1197-11ea-8c1f-
01aa75ed71a1/language-en/format-PDF.
73
George Anthony Gal et al., “Artificial Intelligence in Space.”

437

Electronic copy available at: https://ssrn.com/abstract=3880911

Bratu, Lodder & van der Linden

A. INTERNATIONAL RESPONSIBILITY UNDER THE OUTER

SPACE TREATY
The Outer Space Treaty and the Liability Convention do not expressly
provide that the Liability Convention should serve as the sole remedy for
seeking compensation for damages caused by space objects. Thus, there is no
reason why the more general concept of state responsibility could not be used
as an alternative for obtaining compensation for damage in cases where the
liability concept does not offer a solution.74
The difficulty in applying this solution is related to the grounds based
on which liability and responsibility are attributed. A state is responsible for
“national activities” in outer space under Article VI of the Outer Space Treaty,
while a state is liable pursuant to its capacity as a “launching state” under
Article VII of the Outer Space Treaty and the corresponding Articles II and III
under the Liability Convention.
The challenges triggered by the intersection of these legal provisions may
be more visible in the following example: a telecommunications satellite that
was launched by state X is then sold, while in orbit, to a private operator of
state Y. State Y was not involved in the launch, therefore it cannot be cat-
egorised as a ‘launching State’ of the satellite. If the satellite caused damage
compensable under the Liability Convention, the original launching state, i.e.,
state X, would have remained liable even if it no longer exercised any ju-
risdiction and control over the satellite. The liability of state X would have
been attributed in accordance with Article VII of the Outer Space Treaty and
the corresponding Article II or III of the Liability Convention, depending on
geographical location of the damage. In the legal doctrine, this was described
as “once a launching state, always a liable one”,75 meaning that the qualifica-
tion of a state as a launching state is directly connected with the attribution of
liability. At the same time, under Article VI of the Outer Space Treaty, state Y
could also be held responsible, but this time, based on the national activity of
that state in the outer space.
The overlap between Article VI and Article VII of the Outer Space Treaty
stands in the way of an effective framework for dealing with incidents re-
sulting in damages caused by space-related activities. Until such overlap is
resolved in a formal manner (e.g., by an authoritative document with the nec-
essary legal force), the confusion remains.
Despite the outlined potential difficulties, Article VI of the Outer Space

74
von der Dunk and Tronchetti, “International Space Law,” 51 – 52.
75
Ibid.

438

Electronic copy available at: https://ssrn.com/abstract=3880911

 Autonomous Space Objects and International Space Law

Treaty offers a theoretical possibility for obtaining compensation in case of

damages caused by space objects, which should be taken into account as an
alternative to the liability gap caused by the insufficient clarity of the liability
regimes introduced by the Liability Convention.

B. AMENDMENTS TO THE LIABILITY CONVENTION

The developments of autonomous technologies, together with ample re-
sources in outer space, provide a strong incentive for more challenging space
activities. Space activities are becoming significantly more complex than at
the time when the Liability Convention was drafted.76 During the negotia-
tions for the Liability Convention, the parties involved acknowledged that it
may eventually prove desirable to have a separate additional treaty when the
presence of human beings in space becomes frequent and numerous.77 Given
that the circumstances present at the moment of its adoption have changed,
it might be the time to have a second look at the provisions of the Liability
Convention.
The analysis of Article II and Article III of the Liability Convention re-
vealed that attributing liability for damages caused by autonomous space ob-
jects may be hindered by the lack of clarity in what concerns the fundamental
concepts included in the Liability Convention (more specifically: negligence
and intention) and the basis for attributing liability. The optimal solution for
mitigating these challenges would be an amendment of the Liability Conven-
tion.
However, the process of amending the Liability Convention may prove to
be extremely complicated. The initial drafting and negotiation of the Liability
Convention required an extended period of time. The entire process was initi-
ated in 1962, while the final version of the Convention was made available in
1971 and it proved to be the result of one of the most difficult treaty negotia-
tions since 1945.78 Establishing this regime encountered difficulties due to the
very different approach of the states on various aspects to be included in the
text of the Convention, such as: the determination of the applicable law, in
particular to deal with the amount of damages, the question of whether or not a
limit should be placed on liability incurred under the Convention, the method
for settlement of disputes and the character of the decision rendered, and the
status of international organizations with regard to the Convention.79
76
Zhao, “The 1972 Liability Convention”.
77
Herbert Reis, “Some Reflections on the Liability Convention for Outer Space,” Journal of
Space Law 6, no. 2 (1978): 125–28.
78
Ibid.
79
Valérie Kayser, Launching Space Objects, 33.

439

Electronic copy available at: https://ssrn.com/abstract=3880911

Bratu, Lodder & van der Linden

The discussion surrounding potential amendments of the Liability Con-

vention have been taken place ever since its enactment. However, because of
the political situation (Cold War) during the drafting period, scholars tend to
be satisfied with the achievements made.80
One of the pressing needs requiring special attention is the definition of the
“space object”, due to the practical difficulties raised by the current provisions
of the Liability Convention, as already outlined under Section 4.1 above. The
definition of “damages” would also require a more detailed approach. Cur-
rently, the Liability Convention only provides for the availability of damages,
without further clarification concerning direct and indirect damages. Conse-
quently, direct damages should be compensated, but indirect damage still con-
stitute a gray area for the Convention. 81 Another important amendment to the
Liability Convention could refer to the notion of “fault”. By including a defi-
nition of this notion, the Convention would anticipate any future debates con-
cerning the attribution of the fault-liability system in case of damage caused
by an autonomous space object. Another essential amendment could refer to
the intersections between the provisions of the Outer Space Treaty, more spe-
cifically Article VI and VII and those of the Liability Convention, Articles II
and III. The Liability Convention could expressly state that the Convention
itself represents the sole remedy to be used for dealing with compensation for
damages caused by space objects, with the explicit exclusion of Article VI of
the Outer Space Treaty as an option for a remedy. Such an amendment would
lay the grounds for a transparent and efficient mechanism for attributing liabil-
ity and seeking compensation for damages caused by space-related activities.
In what concerns the legal basis for initiating a potential amendment, there
seems to be agreement between space lawyers, who are of the opinion that
potential amendments can be initiated in accordance with article 25 of the Li-
ability Convention, which states that any State Party to this Convention may
propose amendments to this Convention. Amendments shall enter into force
for each State Party to the Convention accepting the amendments upon their
acceptance by a majority of the States Parties to the Convention and thereafter
for each remaining State Party to the Convention on the date of acceptance
by it.82
Nevertheless, until either an official amendment process is initiated, or an
equally authoritative decision of an international court or tribunal would rule
in a given dispute providing clarification of the liability regimes, the existing
80
Zhao, “The 1972 Liability Convention”.
81
Ibid.
82
Edward R. Finch, “Outer Space Liability: Past, Present and Future,” The International
Lawyer 14, no. 1 (1980): 123–127.

440

Electronic copy available at: https://ssrn.com/abstract=3880911

 Autonomous Space Objects and International Space Law

provisions of the Liability Convention remain effective.

C. APPLYING PRINCIPLES OF INTERNATIONAL LAW

In any legal system, there is the possibility of gaps and silences.83 There-
fore, assuming that neither the Liability Convention nor the Outer Space Trea-
ty can offer a solution for obtaining compensation in case of damages caused
by space objects and that the amendment of the Liability Convention does not
represent an efficient solution, then, states can make use of general principles
of international law.
For example, the principles from several international law cases concern-
ing fault standards and due diligence obligations can be applied to space-relat-
ed activities, in an attempt to identify solutions for attributing fault-based lia-
bility. The due diligence obligations of a state were defined in the International
Court of Justice’s (ICJ) decision Corfu Channel.84 According to this decision,
states are obliged not to knowingly allow their territory to be used for acts
contrary to the rights of other States.85 Under this due diligence obligation, the
relevant fault standard is so-called ‘constructive knowledge’,86 meaning that
a state is expected to be aware of certain important activities being developed
on its territory. In this specific case, Albania did not necessarily know of the
presence of mines in its territorial waters but it ‘should have known’.
Applying this obligation in the outer space context would mean that, by
virtue of a launching state’s control over its activities, it is presumed that a
state will have constructive knowledge about the circumstances surrounding
the operation of the space object, including, for example, the possibility of
its collision with another space object.87 As such, irrespective of the poten-
tially unforeseeable behaviour of a space object equipped with AI capabilities,
launching states can be expected to know about the circumstances relevant to
their active space objects’ operation. In particular, a launching state should be
83
Prosper Weil, “The Court Cannot Conclude Definitively . . . Non Liquet Revisited Chapter
1: Questions of Theory,” Columbia Journal of Transnational Law 36, no. Issues 1 & 2 (1998):
109–119.
84
International Court of Justice, “Latest Developments | Corfu Channel (United Kingdom
of Great Britain and Northern Ireland v. Albania) | International Court of Justice,” accessed
February 14, 2021, https://www.icj-cij.org/en/case/1.
85
William W. Bishop, J. G. Guerrero, and E. Hambro, “The Corfu Channel Case (Merits),”
American Journal of International Law 43, no. 3 (July 1949): 558–89, https://doi.
org/10.2307/2193658.
86
Frauke Renz, State responsibility and new trends in the privatization of warfare (Edward
Elgar Publishing Limited 2020), 86
87
Joel A Dennerley, “State Liability for Space Object Collisions: The Proper Interpretation of
‘Fault’ for the Purposes of International Space Law,” European Journal of International Law
29, no. 1 (May 8, 2018): 281–301, https://doi.org/10.1093/ejil/chy003.

441

Electronic copy available at: https://ssrn.com/abstract=3880911

Bratu, Lodder & van der Linden

aware of the risk that it takes by launching an autonomous space object. This
knowingly taking a risk would then be the justification for assigning liability
to that state should the risk materialise in the sense that the autonomous object
causes damage.
In a distinct case concerning Application of the Convention on the Preven-
tion and Punishment of the Crime of Genocide (Bosnia and Herzegovina v.
Serbia and Montenegro)88 the ICJ extended the application of due diligence
under the Corfu Channel case. This means that the due diligence obligation is
not exclusively connected to a state’s control over its territory. The due dili-
gence obligation also covers elements under a state’s jurisdiction and control
that it has power over or has the capacity to influence.89 In an outer space con-
text, the space objects’ launch and operation are presumed to be activities that
launching states have control over. This suggests that the best-efforts obliga-
tion of due diligence to prevent acts, such as causing space object collisions,
that would cause damage to another state is a duty incumbent on launching
states.90
The launching state’s responsibility also applies in cases where it is not the
state itself that is involved in the launching, but a New Space private party – as
there is not yet in the Treaties a provision to hold a private company liable for
damages caused in space.

VI. CONCLUSION
In recent years, the space industry has been revolutionised. Increas-
ing budgets laid the grounds for technological advancements. Space objects
launched by states and private actors are becoming increasingly sophisticated
given their AI capabilities. The autonomy of space objects has become a prior-
ity for states as well as for private actors. Autonomous space objects are used,
among others, to monitor the operation of satellites and climate change, are
implemented in space stations operations as virtual assistants for astronauts
and they support exploration on planets, where on-site conditions are still too
dangerous for humans. Moreover, there is an increasing number of situations
in which human control over a space object is no longer economically or prac-
tically feasible.

88
International Court of Justice, “Latest Developments: Application of the Convention on the
Prevention and Punishment of the Crime of Genocide (Bosnia and Herzegovina v. Serbia and
Montenegro),” accessed February 14, 2021, https://www.icj-cij.org/en/case/91.
89
Dennerley, “State Liability for Space Object Collisions,” 281.
90
Ibid.

442

Electronic copy available at: https://ssrn.com/abstract=3880911

 Autonomous Space Objects and International Space Law

The introduction of AI systems in space-related activities does not come

without legal implications, especially regarding liability matters. Autonomous
space objects may display unforeseeable behaviour by being capable of tak-
ing decisions on their own, thus challenging the existing liability frameworks,
under the Liability Convention. The optimal solution for mitigating such chal-
lenges would be the amendment of the Liability Convention for addressing
the latest technological advancements. However, the process of amending the
Liability Convention may prove to be extremely complicated. The negotiation
process prior to the entering into force of the Convention took a long time due
to conflicting interests of the parties involved. Initiating an amendment pro-
cess for the Liability Convention should take in consideration these potential
drawbacks. When the Liability Convention is unable to offer a solution for
attributing liability, there is no fundamental reason why the general responsi-
bility framework, as provided by the Article VI, under the Outer Space Treaty
could not be invoked. A potential solution for situations in which the Liability
Convention is unable to offer a solution would be applying principles of inter-
national law, as an alternative fall-back mechanism.
Initiating legal discussions related to the potential damages caused by au-
tonomous space objects is required in order to avoid situations where liability
cannot be attributed: a liability gap. Mapping the legal framework in advance
is preferable to a post-factum intervention, in which scenario an incident
would have occurred already.

443

Electronic copy available at: https://ssrn.com/abstract=3880911

Bratu, Lodder & van der Linden

BIBLIOGRAPHY
Articles in journals and periodicals
Bishop, William W. J., G. Guerrero, and E. Hambro. “The Corfu Channel Case (Mer-
its).” American Journal of International Law 43, no. 3 (1949): 558–89.
Cheng, Bin. “International Responsibility and Liability for Launch Activities.” Air
and Space Law 20, no. 6 (December 1, 1995): 297-310.
Chien, Steve, and Kiri L. Wagstaff. “Robotic Space Exploration Agents,” in Science
Robotics 2, no. 7 (2017): 1-2.
Chien, Steve, et. al. “The Future of AI in Space.” IEEE Intelligent Systems 21, no. 04
(July 1, 2006): 64–69.
Christol, Carl Q. “International Liability for Damage Caused by Space Objects.” Amer-
ican Journal Of International Law 74, no. 2 (1980): 346-371.
Dennerley, Joel A. “State Liability for Space Object Collisions: The Proper Interpreta-
tion of ‘Fault’ for the Purposes of International Space Law.” European Journal of
International Law 29, no. 1 (May 8, 2018): 281–301.
Finch, Edward R. “Outer Space Liability: Past, Present and Future.” The International
Lawyer 14, no. 1 (1980): 123–127.
Firestone, Marc S. “Comment: Problems in the Resolution of Disputes Concerning
Damage in Outer Space.” Tulane Law Review, (January 1985): 747-772.
Gorove, Stephen. “Liability in Space Law: An Overview Space Law.” Annals of Air
and Space Law 8, (1983): 373–380.
Gunkel, David J. “The Other Question: Can and Should Robots Have Rights?” Ethics
and Information Technology 20, no. 2 (2017): pp. 87-99.
Lee, Yoon. “Registration of Space Objects: ESA Member States’ Practice.” Space
Policy 22, no. 1 (2006): 42–51.
Martin, Anne-Sophie and Steven Freeland “The Advent of Artificial Intelligence in
Space Activities: New Legal Challenges,” Space Policy 55 (February 1, 2021):
101408.
Qolomany, Basheer, et al. “Leveraging Machine Learning and Big Data for Smart
Buildings: A Comprehensive Survei.” IEEE Access 7 (2019): 90316–90356.
Reis, Herbert.“Some Reflections on the Liability Convention for Outer Space.” Jour-
nal of Space Law 6, no. 2 (1978): 125–128.
Steve Chien and Kiri L. Wagstaff, “Robotic Space Exploration Agents,” in Science
Robotics 2, no. 7 (2017): 1-2.
Weil, Prosper. “The Court Cannot Conclude Definitively . . . Non Liquet Revisited
Chapter 1: Questions of Theory.” Columbia Journal of Transnational Law 36,
no. Issues 1 & 2 (1998): 109–119.
Zhao, Yun. “The 1972 Liability Convention: Time for Revision?” Space Policy 20,
no. 2 (2004): 117–122.

Books and book chapters

Vernile, Alessandra. The Rise of Private Actors in the Space Sector. New York:
Springer, 2018.
Legal Documents
Treaty on Principles Governing the Activities of States in the Exploration and Use

444

Electronic copy available at: https://ssrn.com/abstract=3880911

 Autonomous Space Objects and International Space Law

of Outer Space, Including the Moon and Other Celestial Bodies. 610 UNTS 205
(opened for signature 27 January 1967, entered into Force 10 October 1967).
United Nations Convention on Registration of Objects Launched into Outer Space.
1023 UNTS 15 (opened for signature 12 November 1975, entered into force on
15 September 1976).
United Nations Convention on the International Liability for Damage Caused by
Space Objects. 961 UNTS 13810 (opened for signature on 29 March 1972, en-
tered into force 20 March 1975).

Web sources
“Floating Robot Cimon sent to International Space Station.” BBC News, 29 June 2018,
accessed 12 February 2021, https://www.bbc.com/news/technology-44655675.
“What Is Artificial Intelligence and How Is It Used?” News European Parlia-
ment, 9 April 2020, https://www.europarl.europa.eu/news/en/headlines/
society/20200827STO85804/what-is-artificial-intelligence-and-how-is-it-used.
Airbus. “Astronaut Assistant CIMON-2 Is on its way to the International Space Sta-
tion.” Accessed 12 February 2021, https://www.airbus.com/newsroom/press-re-
leases/en/2019/12/astronaut-assistant-cimon2-is-on-its-way-to-the-international-
space-station.html.
Berger, Eric. “Meet Ravn X—A fully autonomous, air-launched rocket for small sat-
ellites.” Ars Technica, accessed 20 December 2020, https://arstechnica.com/sci-
ence/2020/12/meet-ravn-x-a-fully-autonomous-air-launched-rocket-for-small-
satellites/.
Berquand, Audrey and Deep Bandivadekar. “Five Ways Artificial Intelligence Can
Help Space Exploration,” The Conversation, 25 January 2021, accessed February
12, 2021, http://theconversation.com/five-ways-artificial-intelligence-can-help-
space-exploration-153664.
Cambridge Dictionary. “Gross Negligence.” Accessed 20 December 2020, https://dic-
tionary.cambridge.org/dictionary/english/gross-negligence.
European Parliament Think Tank. “How Artificial Intelligence Works.” Accessed
14 February 2021, https://www.europarl.europa.eu/thinktank/en/document.
html?reference=EPRS_BRI(2019)634420.
European Space Agency. “Automating Collision Avoidance.” accessed February 12,
2021, https://www.esa.int/Safety_Security/Space_Debris/Automating_collision_
avoidance.
European Space Agency. “Robots in Space.” Accessed 12 February 2021, https://
www.esa.int/Enabling_Support/Preparing_for_the_Future/Discovery_and_Prep-
aration/Robots_in_space2
Farias, Humberto. “Machine Learning Vs Predictive Analytics: What’s the Differ-
ence? Data Science.” Accessed 11 February 2021, https://www.conceptatech.
com/blog/machine-learning-vs-predictive-analytics-what-is-the-difference.
Gal, George Anthony, Cristiana Santos, Lucien Rapp, Réka Markovich, Leendert van
der Torre. “Artificial intelligence in space.” available at https://www.research-
gate.net/publication/342377395_Artificial_intelligence_in_space.
Information Commissioner’s Office. “Big Data, Artificial Intelligence, Machine
Learning and Data Protection.” Accessed https://ico.org.uk/media/for-organisa-
tions/documents/2013559/big-data-ai-ml-and-data-protection.pdf.

445

Electronic copy available at: https://ssrn.com/abstract=3880911

Bratu, Lodder & van der Linden

International Court of Justice. “Latest Developments: Application of the Convention

on the Prevention and Punishment of the Crime of Genocide (Bosnia and Her-
zegovina v. Serbia and Montenegro.” Accessed 14 February 2021, https://www.
icj-cij.org/en/case/91.
International Court of Justice. “Latest Developments: Corfu Channel (United King-
dom of Great Britain and Northern Ireland v. Albania).” Accessed 14 February
2021, https://www.icj-cij.org/en/case/1.
Jet Propulsion Laboratory. “NASA’s Mars Rover Drivers Need Your Help.” Accessed
12 February 2021, https://www.jpl.nasa.gov/news/nasas-mars-rover-drivers-
need-your-help/.
Jet Propulsion Laboratory. “NASA’s Mars Rover Drivers Need Your Help.” Accessed
14 February 2021, https://www.jpl.nasa.gov/news/news.php?feature=7675.
Macaulay, Thomas. “AI to Help World’s First Removal of Space Debris.” Neural,
30 October 2020, accessed 12 February 2021, https://thenextweb.com/neu-
ral/2020/10/30/ai-to-help-worlds-first-removal-of-space-debris.
Oberhaus, Daniel. “How NASA Built a Self-driving Car for Its Next Mars Mission.”
Wired, 21 July 2020, accessed 12 February 2021, https://www.wired.com/story/
how-nasa-built-a-self-driving-car-for-its-next-mars-mission/.
Schmelzer, Ron. “How is AI Helping to Commercialize Space?” Forbes, 21 March
2020, accessed February 12, 2021, https://www.forbes.com/sites/cognitive-
world/2020/03/21/how-is-ai-helping-to-commercialize-space.

Others
Samoli, Sofia, et.al. “AI Watch: Defining Artificial Intelligence.” (Luxembourg: Pub-
lications Office of the European Union, 2020).
von der Dunk, Frans. “Liability versus Responsibility in Space Law: Misconception
or Misconstruction?” In Proceedings of the 34th Colloquium on the Law of Outer
Space, 363-371, 1992.
Kopal, Vladimir. “Some Remarks on Issues Relating to Legal Definitions of Space
Objects, Space Debris and Astronaut.” In Proceedings of the 37th Colloquium on
the Law of Outer Space, 99-106, 1994.
European Commission Brussels. “Communication from the Commission to the Eu-
ropean Parliament, the European Council, the European Economic and Social
Committee and the Committee of the Regions, Coordinated Plan on Artificial In-
telligence.” 7 December 2018, https://eur-lex.europa.eu/legal-content/EN/TXT/
PDF/?uri=CELEX:52018DC0795&from=EN.
European Space Policy Institute. “ESPI Yearbook 2019: Space Policies, Issues
and Trends.” https://espi.or.at/?view=article&id=468:espi-yearbook-
2019&catid=29.
von der Dunk, Frans. “Too-Close Encounters of the Third Party Kind: Will the Liabil-
ity Convention Stand the Test of the Cosmos 2251-Iridium 33 Collision?” Space,
Cyber, and Telecommunications Law Program Faculty Publications, January 1,
2010, https://digitalcommons.unl.edu/spacelaw/28.
Mathieu, Pierre-Philippe, Sveinung Loekken, et. al., “Towards a European AI for
Earth Observation Research & Innovation Agenda,” in Workshop at ESA Φ-lab
(European Space Agency, 2018), 1-20, https://blogs.esa.int/philab/files/2018/07/
Towards-a-European-AI-for-Earth-Observation-Research-Innovation-Agenda-.

446

Electronic copy available at: https://ssrn.com/abstract=3880911

 Autonomous Space Objects and International Space Law

pdf.
European Commission. “A Definition of Artificial Intelligence: Main Capabilities and
Scientific Disciplines,” in Independent High-Level Expert Group on Artificial In-
telligence set up by the European Commission, 8 April 2019.

447

Electronic copy available at: https://ssrn.com/abstract=3880911